Risk Factors for Stress Ulcers
Patients considered to be high risk for stress ulcers:
- > 48 hours mechanical ventilation*
- Severe head injury
- Severe trauma
- > 30% BSA burns
- Emergent or major surgery
- Hypotension, shock, severe sepsis, or MOF
- Renal or hepatic failure
- Ulcerogenic pharmacologic agents
- History of bleeding ulcers
Two independent risk factors for bleeding were identified in that study: acute respiratory failure requiring mechanical ventilation for more than 48 h and coagulopathy [NEJM 330: 377, 1994]. For the rest, one must have at least two concurrent risk factors to be considered high risk. The same group later found that in mechanically ventilated patients, acute renal failure is an independent risk factor [Cook D et al. CCM 27: 2812, 1999]
Preservation of gastric blood flow is essential but difficult to monitor – sublingual capnometry is promising but unproven [Chest 120: 923, 2001]. Enteric feedings reduce the risk of stress ulcer bleeding while providing nutrients [Chest 90: 2, 1986] and should be used if possible – enteral feeding has been associated with a reduced risk of stress gastritis [Cook D et al. CCM 27: 2812, 1999].
Available Pharmacologic Agents
Pharmacologic agents include sucralfate, H2-blockers, and PPIs. Sucralfate is least expensive, has been proven to reduce overt bleeding from ICU patients [JAMA 275: 308, 1996], but has important interactions with warfarin, digoxin, fluoroquinolones, and phenytoin among others (give these drugs 2 hours before sucralfate, which is given 1g q6h). H2-blockers don’t have as many interactions but must be renally dosed. They have also been shown to reduce clinically significant bleeding [JAMA 275: 308, 1996], but at the cost of increased incidence of pneumonia [JAMA 292: 1955, 2004]. PPI’s are preferred agents for GERD and PUD because they offer more complete acid reduction and also do not lead to tolerance (H2-blockers do). PPIs are inactivated by acid, so should probably be given IV in the ICU (pantoprazole).
Sucralfate vs. Ranitidine
Because of its acid-sparing effect, sucralfate was thought by some to be superior to H2-blockers, however a multicenter, randomized, blinded, placebo-controlled trial between sucralfate and ranitidine in 1200 patients who required mechanical ventilation showed a significant reduction in clinically important gastrointestinal bleeding (3.8% vs. 1.7% p=0.02) in patients on ranitidine, but no significant difference in ventilator-associated pneumonia (p = 0.19), and no significant difference in ICU mortality [NEJM 338: 791, 1998]
A small RCT of 94 patients with ICH within 7 days were randomized to ranitidine vs. sucralfate and conservatively managed. GH occurred in 11 (23.4%) patients in placebo, 5 (11.1%) in ranitidine and 7 (14.3%) in sucralfate group, which was not significant. There were 13 (27.7%) deaths in the placebo group, 5 (11.1%) in the ranitidine group, and 12 (24.5%) in the sucralfate group. Pneumonia occurred in 5 of the placebo group (10.6%), 2 of the ranitidine group (4.4%) and 5 of the sucralfate group (10.2%), which was not significantly different. The authors concluded that neither ranitidine nor sucralfate seem to significantly prevent GH or reduce 1-month mortality [J Neurol Sci 239: 5, 2005]
Although not a “drug” per se, the use of enteral feeds is clearly associated with a reduced incidence of stress ulceration
Meta-Analyses of Ranitidine
There are two remaining controversies in the discussion of pharmacologic prophylaxis – first, whether or not all ICU patients require prophylaxis, and second, which agent is optimal.
A recent meta-analysis suggested that ranitidine is ineffective in the prevention of GI bleeding in ICU patients and might increase the risk of pneumonia, but acknowledged that the findings were based on small numbers of patients, and thus firm conclusions could not be proposed. It also found that mortality rates between sucralfate and H2-blockers were statistically the same [BMJ 321: 1103, 2000]. A French retrospective analysis of over 1400 ICU patients suggested that stress-ulcer prophylaxis does not influence the clinically significant gastrointestinal bleeding rate in ICU patients [Intensive Care Med 29:1306, 2003]
A previous meta-analysis concluded that all H2-blockers are equally effective [JAMA 275: 308, 1996]. These drugs are usually given IV, although they can be given PO. Continuous infusion is more effective than bolus [Ann Int Med 121: 568, 1994]. Some data have suggested that pantoprazole IV may be more effective than cimetidine in preventing stress ulcers [Gastroenterology 120 suppl 1:A-157, abstract 838, 2001], although it is important to note that other data have shown no difference in other types of GI bleeding [Gastroenterology 116: 4, 1999]. According to Marino, PPI’s are not as well studied in the ICU setting, but given the effectiveness of sucralfate and H2-blockers as well as their extreme ability to reduce acid production, their use is probably unnecessary [Marino]
Potential Adverse Effects of Empiric GI Prophylaxis
Some intensivists hypothesize that “de-acidifying” the stomach adversely affects the ecology of the gastrointestinal tract (the low pH helps maintain gut sterility) and, indeed, H2-blockers have been shown in some studies to increase the incidence of pneumonia. Laheij et al. examined over 5000 cases of community-acquired pneumonia (CAP) and showed a clear relationship between the risk of CAP and use of H2-blockers (RR 1.63) and PPIs (RR 1.89) [Laheij RJ et al. JAMA 292: 1955, 2004]. Driks et al. compared H2-blockers to sucralfate in 130 mechanically ventilated patients and found a 2-fold increase in the incidence of VAP in those given sucralfate [Driks MR et al. N Engl J Med 317: 1376, 1987]. Cook’s meta-analysis stated that “Prophylaxis with sucralfate is associated with a lower rate of nosocomial pneumonia and mortality” [Cook DJ. Hepatogastroenterology 30(suppl 210): 48, 1995]
That said, a multicenter, randomized, blinded, placebo-controlled trial between sucralfate and ranitidine in 1200 patients who required mechanical ventilation showed no significant difference in ventilator-associated pneumonia (p = 0.19), and no significant difference in ICU mortality [NEJM 338: 791, 1998]. Kantorova et al found no difference in over 200 patients randomized to various regimens of prophylaxis (including placebo) [Kantorova H et al. Hepatogastroenterology 51: 757, 2004]. The issue remains unsettled.
Antacids clearly increase the risk of clostridium difficile infection in the community [Dial S et al. JAMA 294: 2989, 2005]. This association has not been shown in critically ill patients.
There is likely an association between H2-blocker use and thrombocytopenia [Ecker RD et al. J Neurosurg Anesthesiol 16: 291, 2004] although this is controversial [Wade EE et al. Intensive Care Med 28: 459, 2002]. There are case reports of an association between PPI use and thrombocytopenia [Watson TD et al. Ann Pharmacother 40: 758, 2006] but not the large associative data seen with H2-blockers. Note that it is impossible to separate causality from association in retrospective studies.